flashseq_demo_human
Run Summary
| Summary | Value |
|---|---|
| Run ID | flashseq_demo_human |
| Genome Assembly | homo_sapiens_GRCh38.104 |
| # Plates Detected | 1 |
| # Groups | 1 |
| # Total Wells Processed | 384 |
| # Excluded Wells (empty FASTQs) | 155 |
| # Cells (Before QC) | 229 |
| # Cells (After QC) | 191 |
| % Cells Filtered Out | 16.59 % |
| Seurat Version | 5.1.0 |
| R Version | R version 4.2.3 (2023-03-15) |
| Analysis Date | 2025-09-18 |
Introduction
This file contains an overview of the results from flashseq_demo_human.
Whenever possible we compare the reference group to the other groups using a two-sided Wilcoxon rank sum test. Only P-values inferior to 0.05 are displayed and encoded as follows:
- p-value < 0.05: “*”
- p-value < 0.01: “**”
- p-value < 0.001: “***”
- p-value < 0.0001: “****”
Raw Reads
Number of raw reads per cell.
A uniform distribution of the reads between cells is expected when following the standard FLASH-seq protocol.
Raw Reads - Plate View
Drop outs originating from FACS sorting issues are often characterized by a low number of associated reads.
Mapping Stats
From STAR mapping statistics.
Uniquely Mapped reads: Reads associated to a single position in the reference genome. Multimapped reads: Reads mapping at multiple positions onto the reference. Unmapped reads: Reads that do not map to the reference genome or present a too high number of multimapped positions.
The percentage of uniquely mapped reads can vary between cell types. Good quality cells typically display values >70%.
Sequencing Depth Saturation
Read Distribution
The read distribution between 3’/5’UTR, exonic coding sequence, intronic and intergenic regions are estimated using ReSQC. Read tag percentages are displayed.
The majority of the reads will be distributed between CDS exon and intronic categories.
Mitochondrial Genes
Gene-body Coverage
Distribution of the reads along the gene-body. As a full length protocol, FLASH-seq produces uniform distribution from 5’ to 3’.
Estimated using ReSQC.
Population Overview
First glance at the single-cell populations. This analysis is provided for reference only. Due to its automated nature, we highly recommend fine-tuning the parameters before annotating your cell populations.
Filtering
Cells are considered as outliers and filtered out of the population analysis if they don’t meet the following criteria::
- Mitochondria reads: 0.1%-10%
- Number of detected genes: >1000 and <(median + 2x median absolte deviation)
- Number of associated reads (exonic): median +/- 3x median absolte deviation
Clustering
Cell populations are identified following the standard Seurat (V5) pipeline. Briefly, the data are log-normalized and then rescaled, regressing out the read counts. The top 2000 most variable features are extracted using ‘vst’ function and used to run a principal component analysis. The number of dimensions is chosen as follows:
- Number of principal component (PC) that explain >80% of the variability
- Number of principal component where the difference between two consecutive PC is <0.25%
Both metrics are compared and the smallest number of PC is selected. These dimensions are used to cluster cells into groups (shared nearest neighbor) and to display the results (UMAP).
Feature Distribution
Markers
Wilcoxon-test - LogFC > |1| - minimal population expression > 0.5 - Bonferroni correction for multiple testing.
Only genes with adjusted p-values <0.05 are displayed.
Failure Rate
Percentage of low quality cells after sequencing. Defined as:
- less than 100K raw reads
- more than 25% unmapped reads
- less than 75K uniquely mapped reads
